All of lore.kernel.org
 help / color / mirror / Atom feed
From: Martin Jansa <martin.jansa@gmail.com>
To: Valentin Popa <valentin.popa@intel.com>
Cc: openembedded-core@lists.openembedded.org
Subject: Re: [PATCH] Mesa: upgrade to 9.2.2
Date: Wed, 22 Jan 2014 13:57:24 +0100	[thread overview]
Message-ID: <20140122125724.GU4100@jama> (raw)
In-Reply-To: <52DFBE2C.2080209@intel.com>

[-- Attachment #1: Type: text/plain, Size: 3892 bytes --]

On Wed, Jan 22, 2014 at 02:48:44PM +0200, Valentin Popa wrote:
> On 01/22/2014 02:28 PM, Valentin Popa wrote:
> > On 01/08/2014 07:38 PM, Martin Jansa wrote:
> >> Hi,
> >>
> >> since this patch:
> >> commit 263e654e5d28fa7f6b0c8cc23f57a31448e88a40
> >> Author: Valentin Popa <valentin.popa@intel.com>
> >> Date:   Wed Oct 30 12:27:39 2013 +0200
> >>
> >>      Mesa: upgrade to 9.2.2
> >>
> >>      * License is still MIT
> >>      * removed patches were already merged or
> >>        another solution was backported.
> >>
> >> I have to define MESA_EGL_NO_X11_HEADERS in qtbase and qtwebkit again to
> >> build them without xlib.h available (DISTRO without x11 in
> >> DISTRO_FEATURES).
> >>
> >> Are you sure that
> >> meta/recipes-graphics/mesa/mesa-9.1.6/0003-EGL-Mutate-NativeDisplayType-depending-on-config.patch 
> >>
> >> was resolved in upstream?
> >>
> >> Regards,
> > Hi,
> > Sorry for this late response,
> > I removed that patch after testing wayland without X and worked.
> > I've tried yesterday to build qtbase  (meta-qt5) to see exactly what's 
> > the problem and determine if indeed part of that patch has to be 
> > upstream, but it was failing for not finding libGL (which of course, 
> > we don't build with mesa when 
> I meant: libGL is not built together with the other mesa libraries when 
> DISTRO_FEATURES_remove="x11" .
> > DISTRO_FEATURES_remove="x11". Just not having X11 in DISTRO_FEATURES 
> > doesn't mean you'll not have it).
> > Can you please give me some info about your config vars or your 
> > environment so I find out what's the problem!? (I'm trying to 
> > reproduce the issue).

Hi,

I don't see how building libGL relates to this problem, which is more
about mesa headers trying to include xlib.h which isn't available in
sysroot for DISTRO without x11 DISTRO_FEATURE.

Here is part of my discussion with Ross about this problem, let me know
if it's enough for you to see what the problem is:

18:33 < JaMa> rburton: do you still maintain mesa in oe-core?
18:47 < rburton> JaMa: i guess :)
18:47 < JaMa> rburton: http://lists.openembedded.org/pipermail/openembedded-core/2014-January/088072.html
18:48 < JaMa> rburton: I know you were working with upstream on this, so maybe you have better info than Valentin
18:54 < RP> JaMa: you'd be better addressing your email to ross as valentin  is away for a few weeks
18:55 < JaMa> ah OK, I've used valentin as author of that commit
20:43 < rburton> JaMa:  oh, *that*.  noted, will re-open that can of worms tomorrow.
20:43 < rburton> the upstream solution was effectively "use pkgconfig"
20:59 < JaMa> hmm checking if .pc really has that flag :)
21:00 < JaMa> rburton: it's there for gl.pc, but not in gles*.pc, but that doesn't mean it has to be there
21:01 < JaMa> I need to check what that failing build was using in qtbase/qtwebkit
21:04 < rburton> JaMa: if you can mail the build log of what failed, that would be great
21:05 < JaMa> I'll have to simulate that with public layers first :)
21:05 < rburton> heh, yes :)
21:07 < JaMa> looking at qtbase/config.tests/x11/opengl/opengl.pro it just doesn't use pkg-config at all to read gl.pc
21:08 < JaMa> so as soon as  -DMESA_EGL_NO_X11_HEADERS was removed from mesa header it can try to load xlib.h
21:09 < rburton> that entire platform thing is mostly arse
21:10 < JaMa> somewhere in ./src/platformsupport/glxconvenience it loads some mesa header which in turn loads xlib.h and fails
21:10 < JaMa> http://patchwork.openembedded.org/patch/60673/ this is basically the same

The conclusion is that if we cannot fix all apps to use pkgconfig when
searching for mesa, can we restore rebased
meta/recipes-graphics/mesa/mesa-9.1.6/0003-EGL-Mutate-NativeDisplayType-depending-on-config.patch
to fix them automatically?

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

  reply	other threads:[~2014-01-22 12:57 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-01-08 17:38 [PATCH] Mesa: upgrade to 9.2.2 Martin Jansa
2014-01-22 12:28 ` Valentin Popa
2014-01-22 12:48   ` Valentin Popa
2014-01-22 12:57     ` Martin Jansa [this message]
2014-01-22 14:52       ` Otavio Salvador
2014-01-22 15:06       ` Valentin Popa
2014-01-22 15:29         ` Burton, Ross
  -- strict thread matches above, loose matches on Subject: below --
2013-10-24 10:32 Valentin Popa
2013-10-24 10:32 ` Valentin Popa
2013-10-24 13:11   ` Burton, Ross
2013-10-24 13:28     ` Valentin Popa
2013-10-24 13:30       ` Burton, Ross
2013-10-24 13:35         ` Valentin Popa
2013-10-25 12:00         ` Valentin Popa
2013-10-29 17:30 ` Saul Wold
2013-11-01 11:56   ` Burton, Ross

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20140122125724.GU4100@jama \
    --to=martin.jansa@gmail.com \
    --cc=openembedded-core@lists.openembedded.org \
    --cc=valentin.popa@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.